Acquisitions

T1 weighted image

T1-weighted image

T2 weighted image

T2-weighted image

FLAIR

FLAIR image

T1rho map

T1rho map

T2* weighted image

T2*-weighted image

Anatomical Pipelines

The INC standard anatomical processing pipeline, includes: reorientation, denoising, rigid alignment to common space, coregistration to base anatomical (usually T1w), within-modality averaging, FG/BG parsing, intensity bias correction, brain extraction, intensity scaling, and normalization to standard space(s).

Cost is determined by the number of anatomical contrasts (e.g., T1w, T2w, FLAIRw, etc.).

The BRAINSAutoWorkup pipeline optimizes tissue classification through an iterative framework and produces robust parcellation of brain regions including in multi-site settings. Brain regions are labelled using a multi-atlas, similarity-weighted, majority-vote procedure (joint label fusion) with a set of expert-segmented templates adapted from the Desikan-Killiany atlas. Brain regions include cortical and subcortical regions, separated by hemispheres and tissue type (gray or white matter) where appropriate.

FreeSurfer is a set of software tools for the study of cortical and subcortical anatomy. In the cortical surface stream, the tools construct models of the boundary between white matter and cortical gray matter as well as the pial surface. Once these surfaces are known, an array of anatomical measures becomes possible, including: cortical thickness, surface area.

Voxelwise Maps

Estimated using Sequence Adaptive Multimodal SEGmentation (SAMSEG) implemented distributed with FreeSurfer which robustly segments dozens of brain structures without preprocessing. Works best with T1w images and either FLAIR- or T2*-weighted images. WM voxels where the intensity is brighter (FLAIR) or darker (T2star) than cortical gray matter are marked as anomalous. Output includes a map of anomalous WM regions as well as segmentation of GM, WM, and large subcortical structures.

T2* processing involves estimating values by weighting and combining multiple acquisitions varying in echo time (TE), before coregistering to each participant/session native space and normalizing to template space.

Processing of T1rho images involves estimating values by weighting and combining multiple acquisitions, before coregistering to each participant/session native space and normalizing to template space.

Maps of Jacobian Determinants of deformation matrices provide a measurement of the magnitude of the difference in local, voxelwise volume for each participant/session and a standard neurological template. Jacobian maps requires that normalization to the desired template has been completed, and they are useful as a distortion correction in voxel-based morphometry or as variables in tensor-based morphometry.

Myelin maps are calculated as a standardized ratio of T1w and T2w intensities, that provides an estimate related to the local myelin content in each voxel. Myelin maps require coregistered T1w and T2w images and normalization to a standard space for group analyses.

Anatomical Labeling

Manual identification and labelling of gross anatomical lesions. A first pass through SAMSEG is attempted followed by manual inspection and alteration of lesion boundaries. Consensus among 2 independent lesion mappers is acquired for final lesion maps.

Classification of tissue into N categories based on signal intensity. Typical output includes a 3 category classification and  maximum likelihood labels for CSF, GM, and WM as well as posterior probabilities for each class.

Lobules are parcellated using a Joint Label Fusion approach by normalizing to a set of labelled exemplars from the COBRALab (https://cobralab.ca/about/). Briefly, anatomical images are registered using a pre-registered exemplars in a combined, registration quality metric. After normalization, labels are propogated to individual participant native space using a weighted majority vote, where weights correspond to local intensity similarity between the normalized brain and exemplars.

Parcellation of subnuclei of the thalamus, implemented in Freesurfer. This tool produces a parcellation of the thalamus into 25 different nuclei, using a probabilistic atlas built with histological data. The parcellation is based on structural MRI, either the main T1 scan processed through recon-all, or an additional scan of a different modality, which potentially shows better contrast between the nuclei.

Iglesias JE, Insausti R, Lerma-Usabiaga G, Bocchetta M, Van Leemput K, Greve DN, Van der Kouwe A, Fischl B, Caballero-Gaudes C, Paz-Alonso PM, Alzheimer's Disease Neuroimaging Initiative. A probabilistic atlas of the human thalamic nuclei combining ex vivo MRI and histology. Neuroimage. 2018 Dec 1;183:314-26.

Parcellation of Hippocampal subfields and Amygdalar nuclei, implemented in FreeSurfer. The tool uses a probabilistic atlas built with ultra-high resolution ex vivo MRI data (~0.1 mm isotropic) to produce an automated segmentation of the hippocampal substructures and the nuclei of the amygdala.

Hippocampus: A computational atlas of the hippocampal formation using ex vivo, ultra-high resolution MRI: Application to adaptive segmentation of in vivo MRI. (1) Iglesias, J.E., Augustinack, J.C., Nguyen, K., Player, C.M., Player, A., Wright, M., Roy, N., Frosch, M.P., Mc Kee, A.C., Wald, L.L., Fischl, B., and Van Leemput, K. Neuroimage, 115, July 2015, 117-137.

(2) Amygdala: High-resolution magnetic resonance imaging reveals nuclei of the human amygdala: manual segmentation to automatic atlas. Saygin ZM & Kliemann D (joint 1st authors), Iglesias JE, van der Kouwe AJW, Boyd E, Reuter M, Stevens A, Van Leemput K, Mc Kee A, Frosch MP, Fischl B, Augustinack JC. Neuroimage, 155, July 2017, 370-382. (3) Longitudinal method: Bayesian longitudinal segmentation of hippocampal substructures in brain MRI using subject-specific atlases. Iglesias JE, Van Leemput K, Augustinack J, Insausti R, Fischl B, Reuter M. Neuroimage, 141, November 2016, 542-555.

Parcellation of Brainstem substructures, implemented in FreeSurfer. This tool generates an automated segmentation of four different brainstem structures from the input T1 scan: medulla oblongata, pons, midbrain and superior cerebellar peduncle (SCP). We use a Bayesian segmentation algorithm that relies on a probabilistic atlas of the brainstem (and neighboring brain structures) built upon manual delineations of the structures on interest in 49 scans (10 for the brainstem structures, 39 for the surrounding structures).

Bayesian segmentation of brainstem structures in MRI. Iglesias, J.E., Van Leemput, K., Bhatt, P., Casillas, C., Dutt, S., Schuff, N., Truran-Sacrey, D., Boxer, A., and Fischl, B. NeuroImage, 113, June 2015, 184-195.

Non-standard atlases can be implemented for your data for the INPC team. These must be pre-existing labelled brain images, any manual labelling must be billed separately. Cost covers a joint label fusion procedure to accurately transfer labels to standard template space.

Custom manual labelling of anatomical labelling for each session. Automated procedures that can provide partial mappings are first employed if possible, then manually altered or traced by 2 independent raters.

Custom Addons

Construct a template specific to the participants in the current project, using an iterative, unbiased normalization approach.

Custom processing pipelines for specific researcher needs, please contact us for details and discussion.